Machine Learning Based Misbehavior Detection System for False Data Injection Attack in Internet of Vehicles Using Neighbor Public Transport Vehicle Approach

Abstract

The integration of the Internet of Vehicles (IoV) into the Intelligent Transportation System (ITS) has significantly improved its operations, leading to a reduction in road traffic accidents, efficient traffic control, and a decrease in carbon emissions for a more sustainable environment aligned with the Sustainable Development Goals (SDGs). However, the adoption of IoV networks introduces privacy and security challenges. Although cryptographic techniques such as public-key infrastructure (PKI) proposed by standardization bodies like IEEE and ETSI provide protection against outsider attackers, they fail to address the threat posed by insider attackers. To overcome this limitation, researchers have proposed data-centric machine learning-based misbehavior detection frameworks that focus on identifying and mitigating insider attacks. However, existing approaches primarily rely on the Basic Safety Message (BSM) data received from a single vehicle, which allows attackers to manipulate the BSM data without being detected. In this paper, we present a novel data-centric misbehavior detection framework specifically designed to detect false data injection attacks in IoV networks. Our approach leverages neighboring public transportation vehicles (NPTVs) to enhance the detection capabilities. By incorporating the BSM data from NPTVs, we demonstrate the effectiveness of our proposed framework in different scenarios using deep learning, decision tree, and random forest algorithms. Through extensive evaluation, we achieved precision, recall, F1-Score, and accuracy rates of up to 99%, showcasing the superior performance of our approach.